trainableSegmentation.unsupervised

Class ColorClustering

java.lang.Object

trainableSegmentation.unsupervised.ColorClustering

public class ColorClustering
extends Object

This class contains all the library methods to perform color clustering based on a few color space transformations and the Weka clusterers.

Author:

Josu Salinas and Ignacio Arganda-Carreras

Nested Class Summary

Nested Classes

Modifier and Type	Class and Description
static class	ColorClustering.Channel Names of all channels (features) that can be used.

Constructor Summary

Constructors

Constructor and Description
ColorClustering(ij.ImagePlus image) Constructor using only image, features and clusterer will be set as null
ColorClustering(ij.ImagePlus image, int numSamples, ArrayList<ColorClustering.Channel> selectedChannels) Creates color clustering object based on image, number of samples and selected channels.

Method Summary

Modifier and Type	Method and Description
ij.ImagePlus	createClusteredImage(FeatureStackArray featureStackArray) Creates clustered image based on provided FeatureStackArray and using private clusterer, returns as ImagePlus
weka.clusterers.AbstractClusterer	createClusterer(weka.clusterers.AbstractClusterer selectedClusterer) Build clusterer using the current features and the provided number of clusters.
void	createFeatures() Generate the features based on the selected color space channels.
boolean	createFile(String path, weka.core.Instances theInstances) Create ARFF file of features with provided name.
FeatureStackArray	createFSArray(ij.ImagePlus image) Create FeatureStackArray based on provided image and selected channels (private variable)
ij.ImagePlus	createProbabilityMaps(FeatureStackArray featureStackArray) Create probability map based on provided features and current cluster.
weka.core.Instances	getFeaturesInstances() Get features instances that are being used
FeatureStackArray	getFeatureStackArray() Get FeatureStackArray that is being used
ij.ImagePlus	getImage() Get image that is being used
int	getNumSamples() Get number of samples that is being used
weka.clusterers.AbstractClusterer	getTheClusterer() Get abstract clusterer that has been created
boolean	loadClusterer(String path) Load clusterer from path
boolean	saveClusterer(String filename) Save clusterer to a file
void	setChannels(ArrayList<ColorClustering.Channel> selectedChannels) Set channels to use
void	setFeaturesInstances(weka.core.Instances featuresInstances) Set features instances that are to be used
void	setFeatureStackArray(FeatureStackArray featureStackArray) Set FeatureStackArray that is to be used
void	setImage(ij.ImagePlus image) Set image that is to be used
void	setNumSamples(int numSamples) Set number of samples that is to be used
void	setTheClusterer(weka.clusterers.AbstractClusterer theClusterer) Set abstract clusterer to be used

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Detail

ColorClustering

public ColorClustering(ij.ImagePlus image)

Constructor using only image, features and clusterer will be set as null

Parameters:

image - input color image to clusterize

ColorClustering

public ColorClustering(ij.ImagePlus image,
                       int numSamples,
                       ArrayList<ColorClustering.Channel> selectedChannels)

Creates color clustering object based on image, number of samples and selected channels. Creates the features.

Parameters:

image - input color image

numSamples - number of samples to use

selectedChannels - list of color space channels (features) to use

Method Detail

createClusterer

public weka.clusterers.AbstractClusterer createClusterer(weka.clusterers.AbstractClusterer selectedClusterer)

Build clusterer using the current features and the provided number of clusters.

Parameters:

selectedClusterer - clustering Weka model to build

Returns:

built clustering Weka model

createFeatures

public void createFeatures()

Generate the features based on the selected color space channels.

createFSArray

public FeatureStackArray createFSArray(ij.ImagePlus image)

Create FeatureStackArray based on provided image and selected channels (private variable)

Parameters:

image - input color image

Returns:

feature stack array

createProbabilityMaps

public ij.ImagePlus createProbabilityMaps(FeatureStackArray featureStackArray)

Create probability map based on provided features and current cluster.

Parameters:

featureStackArray - input array of feature stacks

Returns:

image containing the probability maps of each cluster

createClusteredImage

public ij.ImagePlus createClusteredImage(FeatureStackArray featureStackArray)

Creates clustered image based on provided FeatureStackArray and using private clusterer, returns as ImagePlus

Parameters:

featureStackArray - input array of feature stacks

Returns:

clusterized image (one label per cluster)

createFile

public boolean createFile(String path,
                          weka.core.Instances theInstances)

Create ARFF file of features with provided name.

Parameters:

path - complete path to output ARFF file

theInstances - dataset of instances to saved to file

Returns:

true if the file was correctly saved, false otherwise

saveClusterer

public boolean saveClusterer(String filename)

Save clusterer to a file

Parameters:

filename - path to file to be saved

Returns:

boolean value showing success/failure

loadClusterer

public boolean loadClusterer(String path)

Load clusterer from path

Parameters:

path - Path of file to be loaded

Returns:

boolean value showing success/failure

getNumSamples

public int getNumSamples()

Get number of samples that is being used

Returns:

number of samples that is being used

setNumSamples

public void setNumSamples(int numSamples)

Set number of samples that is to be used

Parameters:

numSamples - number of samples that is to be used

getImage

public ij.ImagePlus getImage()

Get image that is being used

Returns:

color image being used

setImage

public void setImage(ij.ImagePlus image)

Set image that is to be used

Parameters:

image - color image to be clusterized

getFeatureStackArray

public FeatureStackArray getFeatureStackArray()

Get FeatureStackArray that is being used

Returns:

array of feature stacks used for clustering

setFeatureStackArray

public void setFeatureStackArray(FeatureStackArray featureStackArray)

Set FeatureStackArray that is to be used

Parameters:

featureStackArray - array of feature stacks to use in clustering

getFeaturesInstances

public weka.core.Instances getFeaturesInstances()

Get features instances that are being used

Returns:

set of instances used for building the clusterer

setFeaturesInstances

public void setFeaturesInstances(weka.core.Instances featuresInstances)

Set features instances that are to be used

Parameters:

featuresInstances - set of instances to be used to build the clusterer

getTheClusterer

public weka.clusterers.AbstractClusterer getTheClusterer()

Get abstract clusterer that has been created

Returns:

current clusterer

setTheClusterer

public void setTheClusterer(weka.clusterers.AbstractClusterer theClusterer)

Set abstract clusterer to be used

Parameters:

theClusterer - clustering Weka model to use

setChannels

public void setChannels(ArrayList<ColorClustering.Channel> selectedChannels)

Set channels to use

Parameters:

selectedChannels - list of selected channels